Feedback module 10

Author: msdavidram

learn-pr/wwl/design-alm-process-ai-powered-business-solutions/includes/2-design-alm-process-data-used-ai-models-agents.md

@@ -22,17 +22,17 @@ By the end of this unit, learners are be able to:
 
 #### Treat the following as versioned, promotable ALM artifacts:
 
-**Training & fine-tuning datasets** (raw, curated, feature/embedding sets).
+* **Training & fine-tuning datasets** (raw, curated, feature/embedding sets).
 
-**Evaluation/Testing datasets** and "golden" sets for regression testing.
+* **Evaluation/Testing datasets** and "golden" sets for regression testing.
 
-**Grounding knowledge** (SharePoint/OneDrive corpora, Dataverse tables, wikis, KBs).
+* **Grounding knowledge** (SharePoint/OneDrive corpora, Dataverse tables, wikis, KBs).
 
-**Prompt assets** (system prompts, prompt actions, templates).
+* **Prompt assets** (system prompts, prompt actions, templates).
 
-**Policies & guardrails** (DLP, sensitivity labels, allowed connectors, action limits).
+* **Policies & guardrails** (DLP, sensitivity labels, allowed connectors, action limits).
 
-**Run telemetry** (latency, token/cost, success/failure, safety events) and **feedback**.
+* **Run telemetry** (latency, token/cost, success/failure, safety events) and **feedback**.
 
 ### Environment strategy
 
@@ -158,28 +158,28 @@ The below RACI chart is a representative sample for an organization. It's up to
 
 ## Operational telemetry and drift
 
-Maintain **baseline metrics** per release: latency p95, success %, token/€ per task, safety flags/M runs.
+* Maintain **baseline metrics** per release: latency p95, success %, token/€ per task, safety flags/M runs.
 
-Compare live to baseline; if drift exceeds thresholds, **auto-open an incident**, route to data owner, and pause affected actions.
+* Compare live to baseline; if drift exceeds thresholds, **auto-open an incident**, route to data owner, and pause affected actions.
 
-Rerun evaluation suites nightly/weekly against **golden sets**; store time series for audit.
+* Rerun evaluation suites nightly/weekly against **golden sets**; store time series for audit.
 
 ## Checklists you can paste into your runbooks
 
 ### Go/NoGo before production
 
-[ ]  Data contract approved; asset tagged and discoverable.
-[ ]  Sensitivity labels/DLP rules applied; connectors approved.
-[ ]  Lineage graph current; dataset snapshot **immutable** and versioned.
-[ ]  Evaluation thresholds met; safety risks mitigated.
-[ ]  Residency decision recorded; cross-regions toggle reviewed.
-[ ]  Dashboards, budgets, alerts, and rollback validated in Pre-Prod.
+* [ ]  Data contract approved; asset tagged and discoverable.
+* [ ]  Sensitivity labels/DLP rules applied; connectors approved.
+* [ ]  Lineage graph current; dataset snapshot **immutable** and versioned.
+* [ ]  Evaluation thresholds met; safety risks mitigated.
+* [ ]  Residency decision recorded; cross-regions toggle reviewed.
+* [ ]  Dashboards, budgets, alerts, and rollback validated in Pre-Prod.
 
 ### Retirement
 
-[ ]  Consumers notified; cutover plans executed.
-[ ]  Snapshots archived/deleted per retention; access revoked.
-[ ]  Audit and lineage preserved; catalog updated.
+* [ ]  Consumers notified; cutover plans executed.
+* [ ]  Snapshots archived/deleted per retention; access revoked.
+* [ ]  Audit and lineage preserved; catalog updated.
 
 ## References
 

learn-pr/wwl/design-alm-process-ai-powered-business-solutions/includes/3-design-alm-process-copilot-studio-agents-connectors-actions.md

@@ -8,47 +8,47 @@ Architects learn how to structure solution components into managed lifecycles, e
 
 Copilot Studio solutions typically include:
 
-**Agents** (conversational or autonomous)
+* **Agents** (conversational or autonomous)
 
-**Custom connectors**
+* **Custom connectors**
 
-**Actions, skills, and prompt assets**
+* **Actions, skills, and prompt assets**
 
-**Supporting Dataverse components**
+* **Supporting Dataverse components**
 
-**Environment data and application settings**
+* **Environment data and application settings**
 
 A mature ALM process ensures:
 
-Version control of all solution artifacts
+* Version control of all solution artifacts
 
-Clear separation between development, testing, and production
+* Clear separation between development, testing, and production
 
-Monitoring and validation at each promotion stage
+* Monitoring and validation at each promotion stage
 
-Compliance with governance, DLP, and data residency requirements
+* Compliance with governance, DLP, and data residency requirements
 
-Sustainable lifecycle for updates and deprecations
+* Sustainable lifecycle for updates and deprecations
 
 ## 2. Recommended Environment Strategy
 
 ### Establish at least three core environments:
 
-**Development (Dev)** - building and iterating on agents, connectors, and actions
+* **Development (Dev)** - building and iterating on agents, connectors, and actions
 
-**Test (UAT/QA)** - validating behavior, evaluating prompts, regression checks
+* **Test (UAT/QA)** - validating behavior, evaluating prompts, regression checks
 
-**Production (Prod)** - stable, approved, monitored implementation
+* **Production (Prod)** - stable, approved, monitored implementation
 
 ### Environment design principles:
 
-No direct editing in production
+* No direct editing in production
 
-Enforce role-based access at each tier
+* Enforce role-based access at each tier
 
-Managed solutions only in Test and Prod
+* Managed solutions only in Test and Prod
 
-Use solution layering to isolate changes
+* Use solution layering to isolate changes
 
 ### Professional Visual - Text Diagram
 
@@ -70,49 +70,49 @@ Agents should move through lifecycle stages with predictable governance:
 
 ### 3.1 Development Stage
 
-Draft agent scope, intents, and behaviors
+* Draft agent scope, intents, and behaviors
 
-Build actions and prompts
+* Build actions and prompts
 
-Add knowledge sources in Dev only
+* Add knowledge sources in Dev only
 
-Test agent workflows with edge-case prompts
+* Test agent workflows with edge-case prompts
 
 ### 3.2 Testing Stage
 
-Validate reasoning quality and output patterns
+* Validate reasoning quality and output patterns
 
-Ensure grounding is reliable and compliant
+* Ensure grounding is reliable and compliant
 
-Evaluate event-triggered actions
+* Evaluate event-triggered actions
 
-Run regression tests on all agent topics
+* Run regression tests on all agent topics
 
 ### 3.3 Production Stage
 
-Deploy via managed solutions
+* Deploy via managed solutions
 
-Monitor usage, performance, and safety
+* Monitor usage, performance, and safety
 
-Apply versioning and planned release cycles
+* Apply versioning and planned release cycles
 
-Document change history
+* Document change history
 
 ## 4. ALM for Connectors
 
 Connectors enable Copilot to interact with systems.
 
 ### Connector ALM principles
 
-Build connectors via Dev environment only
+* Build connectors via Dev environment only
 
-Validate authentication flows
+* Validate authentication flows
 
-Apply DLP policies during Test
+* Apply DLP policies during Test
 
-Publish connectors only in Prod after full review
+* Publish connectors only in Prod after full review
 
-Track versions and rollback strategies
+* Track versions and rollback strategies
 
 ### Professional Visual - Connector Release Flow
 
@@ -124,15 +124,15 @@ Actions represent executable steps triggered by agents.
 
 ### Action lifecycle
 
-**Design** - Define outcome, inputs, outputs
+* **Design** - Define outcome, inputs, outputs
 
-**Build** - Implement logic using Power Automate, plugins, or external APIs
+* **Build** - Implement logic using Power Automate, plugins, or external APIs
 
-**Validate** - Confirm correct action sequencing and error handling
+* **Validate** - Confirm correct action sequencing and error handling
 
-**Promote** - Move through Dev → Test → Prod via managed solutions
+* **Promote** - Move through Dev → Test → Prod via managed solutions
 
-**Monitor** - Track success/failure rates and performance metrics
+* **Monitor** - Track success/failure rates and performance metrics
 
 ## 6. Version Control and Release Management
 
@@ -148,65 +148,65 @@ Use Visual Studio Code for Copilot connector development where applicable
 
 ### Adopt release cadence
 
-Monthly or sprint-based release cycles
+* Monthly or sprint-based release cycles
 
-Emergency patch process for critical fixes
+* Emergency patch process for critical fixes
 
-Rollback plan
+* Rollback plan
 
-Change management plan for users
+* Change management plan for users
 
 ## 7. Governance and Compliance Controls
 
 ALM must align with enterprise governance rules:
 
 ### Key controls
 
-DLP policy enforcement
+* DLP policy enforcement
 
-Environment-specific connector rules
+* Environment-specific connector rules
 
-Data residency restrictions
+* Data residency restrictions
 
-Enterprise-approved knowledge sources
+* Enterprise-approved knowledge sources
 
-Review gate for safety, quality, and ethical risk
+* Review gate for safety, quality, and ethical risk
 
-Monitoring dashboards for agent behavior
+* Monitoring dashboards for agent behavior
 
 ### Professional Visual - Governance Checklist
 
-- [✓] DLP Policies Applied
-- [✓] Security Review Completed
-- [✓] Knowledge Sources Verified
-- [✓] Data Residency Confirmed
-- [✓] Risk & Safety Assessment Passed
-- [✓] Monitoring Enabled
+* [✓] DLP Policies Applied
+* [✓] Security Review Completed
+* [✓] Knowledge Sources Verified
+* [✓] Data Residency Confirmed
+* [✓] Risk & Safety Assessment Passed
+* [✓] Monitoring Enabled
 
 ## 8. Monitoring and Continuous Improvement
 
 ### Monitoring signals include:
 
-Agent session success rate
+* Agent session success rate
 
-Action execution health
+* Action execution health
 
-Connector performance
+* Connector performance
 
-Prompt regression—quality changes over time
+* Prompt regression—quality changes over time
 
-User satisfaction insights
+* User satisfaction insights
 
 ### Continuous improvement loop:
 
-Monitor → Analyze → Improve → Release → Validate → Monitor
+* Monitor → Analyze → Improve → Release → Validate → Monitor
 
 ## References
 
-[https://learn.microsoft.com/en-us/microsoft-copilot-studio/authoring-solutions-overview#create-and-manage-solution-pipelines](/microsoft-copilot-studio/authoring-solutions-overview)
+[https://learn.microsoft.com/microsoft-copilot-studio/authoring-solutions-overview#create-and-manage-solution-pipelines](/microsoft-copilot-studio/authoring-solutions-overview)
 
-[https://learn.microsoft.com/en-us/training/paths/extend-microsoft-365-copilot-connectors-visual-studio-code/](/training/paths/extend-microsoft-365-copilot-connectors-visual-studio-code/)
+[https://learn.microsoft.com/training/paths/extend-microsoft-365-copilot-connectors-visual-studio-code/](/training/paths/extend-microsoft-365-copilot-connectors-visual-studio-code/)
 
-[https://learn.microsoft.com/en-us/microsoft-copilot-studio/guidance/architecture/deployment-lifecycle](/microsoft-copilot-studio/guidance/architecture/deployment-lifecycle)
+[https://learn.microsoft.com/microsoft-copilot-studio/guidance/architecture/deployment-lifecycle](/microsoft-copilot-studio/guidance/architecture/deployment-lifecycle)
 
-[https://learn.microsoft.com/en-us/microsoft-copilot-studio/authoring-solutions-overview](/microsoft-copilot-studio/authoring-solutions-overview)
+[https://learn.microsoft.com/microsoft-copilot-studio/authoring-solutions-overview](/microsoft-copilot-studio/authoring-solutions-overview)

learn-pr/wwl/design-alm-process-ai-powered-business-solutions/includes/4-design-alm-process-microsoft-foundry-agents.md

@@ -110,7 +110,7 @@ Foundry agents rely on data retrieved from controlled environments. ALM must enf
 
 * **Use of secure identity management for agent actions**
 
-A centralized governance strategy ensures consistency and reduces operational risk.
+* A centralized governance strategy ensures consistency and reduces operational risk.
 
 ## Deployment and release management